1. WHAT IS INTUITIONAL EVIDENCE?

Before I can begin the argument in earnest, I will need to get precise on what the target of the experimentalists' skepticism is. It is worth considering a paradigmatic example of the practice. In describing the example, I will try as much as possible to remain neutral with respect to any particular theory of intuitional evidence. A standard use of thought experiments involves using them to test various necessity/sufficiency theses. Consider ‘Justified true belief is sufficient for knowledge.’ In judging the correctness of this statement, a philosopher will consider a scenario of the following sort:

Most epistemologists would take an example of this sort as a decisive refutation of the knowledge = justified true belief thesis. Letting ‘p’ express the proposition that there is a zebra on the hill, we may state how this refutation goes through in terms of the following argument:

1. 1. The subject has a justified true belief that there is a zebra on the hill.

2. 2. The subject does not know that there is a zebra on the hill.

3. 3. Therefore, justified true belief is not sufficient for knowledge.

In phrasing the reasoning this way, I do not mean to imply that one must explicitly carry out such reasoning in order to justifiably infer the falsehood of the justified true belief thesis from the thought experiment. But the above argument provides a useful heuristic with which to evaluate the method of thought experimentation.

The premises of the above argument are generally thought of as being known by intuition. By reflecting on the case, one is able to acquire justification for believing that the subject in the case lacks knowledge. Contrasting accounts of the nature of the evidence are possible. On some accounts, the evidence that justifies belief is some kind of intellectual seeming that is distinct from belief; when one has the seeming, one acquires a prima facie justification for believing premises (1) and (2). An account such as Timothy Williamson's denies this: Williamson (Reference Williamson2007) argues that in engaging in the thought experiment, we are able to acquire knowledge of the premises, and this knowledge need not be based on one's psychological states. The details of which account is correct will be passed over, since the experimentalists wish to cast their net wide enough as to raise skepticism about all such accounts.

2. WEINBERG ON EVIDENCE AND HOPEFULNESS

Weinberg's argument begins with his introduction of the terms ‘hopefulness’ and ‘hopelessness.’ Hopefulness and its contrary are properties of evidence sources (perception, memory, instruments, etc. as opposed to the outputs of these sources). Hopelessness involves ‘fallibility unmitigated by a decent capacity for detecting and correcting the errors it entails’ (Weinberg Reference Weinberg2007). A hopeful source of evidence, by contrast, is fallible, but is mitigated by a capacity for detecting and correcting errors. According to Weinberg, perception is a hopeful source of evidence. We are sometimes subject to perceptual illusions, but we have an awareness of what circumstances are liable to make perception more or less reliable, and we know how to correct our perceptual errors. Intuition, by contrast, lacks these benefits.

To make this case, Weinberg sets out four characteristics that contribute to hopefulness; none are individually necessary for hopefulness, but any sources lacking in all four are ‘hopeless’: First is external corroboration: The deliverances of vision can be corroborated by other senses, the deliverances of an electron microscope can be corroborated by other developments in one's theory, etc. Second would be internal coherence: Each of your visual perceptions tends to agree with, or cohere with, your other perceptions. Observations taken with an electron microscope at t tend to agree with observations taken at times subsequent to t. The third contributor to hopefulness is detectability of margins: ‘The practices are sensitive to the conditions in which the device is less likely to give good results’ (Weinberg Reference Weinberg2007: 330). The final measure of hopefulness is theoretical illumination: a working knowledge of how the process or instrument actually functions. We have theoretical illumination with respect to perception because we have a good grasp of the causal processes in play when we use perception; the same is true for electron microscopes. Having set out what hopefulness is, we are in a position to summarize Weinberg's overall argument:

1. 1. (H): Any putative source of evidence that is ‘hopeless’ ought not to be trusted.

2. 2. Philosophers' appeal to intuitions is a ‘hopeless’ practice.

3. 3. Therefore, we ought not to trust philosophers' appeal to intuitions.

For the purposes of the present work, I will grant the second premise of Weinberg's argument. In fact, I am inclined to believe that it is true, given Weinberg's definition of ‘hopelessness.’ My focus will be on Weinberg's principle (H). In justifying (H), Weinberg notes that he cannot appeal to any intuitions, lest his argument be self-defeating. To discharge this burden, Weinberg offers two central considerations in favor of (H): First, there are considerable pragmatic benefits to using hopeful sources of evidence. Most notable of these is that making use of only hopeful sources of evidence makes it far less likely that we commit an error that sets back the progress of some discipline considerably, as the error becomes a part of the edifice upon which other theories are constructed (Weinberg Reference Weinberg2007: 327–8). A strictly pragmatic case for (H) is bound to have certain limitations, however. Jessica Brown puts the point as follows:

Similarly with evidence, Brown argues, the fact that hopeful evidence sources have certain pragmatic benefits does not imply that one should not use ‘hopeless’ sources. It may show that we should prefer hopeful sources to non-hopeful sources, but this does not show that ‘hopeless’ sources may not be used full-stop. There is undoubtedly more to be said about the pragmatic case for (H), but I take it that it is the second part of Weinberg's case which is more instrumental in establishing the truth of (H). For this reason, I will leave the pragmatic case aside and turn to Weinberg's second argument.

3. FROM HOPEFULNESS TO INTUITIONAL SKEPTICISM

Weinberg develops a case for (H) premised on the idea that many of our epistemic practices (journalism, physics, criminal justice, etc.) display a commitment to (H). For instance, journalists are required to have corroborating information for their claims, displaying the importance of external corroboration, an element of hopefulness. The appeals process in jurisprudence illustrates the value of detectability of margins: We must have ways of detecting and correcting for errors (Weinberg Reference Weinberg2007: 333). Weinberg further argues that our best sciences demonstrate a commitment to (H). Drawing from William Bechtel (Reference Bechtel, Creath and Maienschein2000), he observes that biologists were unwilling to endorse electron microscopy until they were reasonably convinced that the results of electron microscopes could be corroborated by other methods, and that their deliverances cohered with extant theories (329). Lastly, the failure of introspectionism as a movement in psychology is best explained, according to Weinberg, by the fact that the practice appealed to sources of evidence that were ‘hopeless’ (333).

Weinberg does not spell out the precise details of how the argument proceeds from the data he cites regarding our epistemic practices to the conclusion that (H) is true. But a reconstruction should not be hard to find. To do so it will be useful to introduce a technical notion of endorsement. Endorse pertains to epistemic practices such as physics, biology, jurisprudence, etc. To endorse p is to treat p as a basis for conduct within that practice. For example, a juror who endorses p is prepared to deliver a verdict that p, a mathematician who endorses p is prepared to treat p as a theorem which may be assumed in the demonstration of other results, etc. Endorsement and belief can come apart: A mathematician may believe Goldbach's conjecture because it has been shown true for however many positive integers, but she will not endorse the conjecture until a formal proof is available.Footnote ¹

I assume that just as there are norms governing belief, there are norms governing endorsement for each of the special practices as well. Violation of these norms leaves one subject to criticism: A journalist who reports information without reliable sources is subject to criticism, a juror who convicts because of an unshakeable hunch is deserving of criticism, etc. We can then rephrase Weinberg's principle (H) as a general norm of endorsement applying across all epistemic practices: One ought not endorse p on the basis of some evidence source E if E is ‘hopeless.’Footnote ²

I have two motivations for moving to talk of endorsement rather than talk of belief. First and most importantly, there is a much tighter connection between the conduct of an epistemic practice and the norms of endorsement implicitly endorsed by those within that practice than there is between that conduct and the relevant norms of belief. A journalist may believe all sorts of things she is not prepared to report in an article: This fact is a much more reliable indicator of the norms of endorsement she appeals to than the norms of belief she holds. She may not necessarily believe that every belief requires corroboration by some trustworthy source. So discussing endorsement rather than belief ensures that the quality of our evidence for (H) or some other hypothesis will be strong.

Secondly, moving to talk of endorsement allows me to avoid taking on certain controversial commitments in the epistemology of belief. My rival norm of endorsement will prominently bring certain pragmatic considerations to the foreground. Proposing this norm as a general rule governing belief could commit me to a potentially problematic pragmatism in the epistemology of belief. Having set all of these considerations out, we can reconstruct Weinberg's argument from the conduct of our epistemic practices to (H) as follows:

1. 1. Our epistemic practices demonstrate a commitment to (H).

2. 2. The commitments of these epistemic practices are a reliable guide to the correct norms for endorsement.

3. 3. Therefore probably, (H) is a correct norm for endorsement.

Premise (2) does not imply the infallibility of current epistemic practice, but it does imply that these practices have a prima facie claim on our support. We can find support for premise (2) in reflecting on the following: Our epistemic practices have a good track record of success, and this is unlikely if they were following bad norms. Premise (2) is not explicitly stated in Weinberg's article, but it is the bridge principle that enables Weinberg to move from empirical facts about our epistemic practice to normative conclusions about which sources of evidence are trustworthy. But I believe that there is adequate support for (2), and that this support need not rely on epistemic intuitions. So if otherwise sound, then Weinberg has succeeded in presenting a challenge to intuitions that avoids epistemic self-defeat.

4. AN ALTERNATIVE TO THE HOPEFULNESS PRINCIPLE

In ‘Intuitions, Evidence and Hopefulness,’ Jessica Brown suggests that the opponent of (H) may accommodate Weinberg's data by appealing to a variety of pragmatic considerations which are particular to the disciplines in question, and do not obviously generalize to all of our epistemic practices. My purpose here is to systematize this suggestion by proposing a rival thesis to (H), and arguing that this norm has greater explanatory power with regard to the conduct of our epistemic practices than does (H): It is able to explain Weinberg's data as well as (H) does, and is also to explain further data that (H) struggles to explain.

Condition (i) attempts to capture the general idea that the more alternatives to some source E are available, the higher the standards E must meet to qualify as an acceptable source. By contrast, the fewer alternatives we have, the lower the standards. Federal exceptions to the prohibitions on hearsay testimony for witnesses ruled unavailable may be an example of the significance of this consideration: Ordinarily, hearsay testimony is inadmissible, but if the witness is unavailable and certain other conditions are met, the testimony may be admitted in accordance with Rule 804 of the Federal Rules of Evidence. This exemption is presumably a result of the fact that there is no alternative way of gaining access to that particular witness testimony.

What sorts of considerations might be relevant to the selection of an evidence source? They vary from discipline to discipline, so nothing like a comprehensive list is possible. Nonetheless, some examples may help to elucidate the general kinds of non-epistemic considerations that might be relevant to the viability of any given putative evidence source. The first set of considerations relevant to the choice of evidence sources within a discipline involves the moral implications of its use. To take an example from jurisprudence, an item of evidence may indicate with near-certainty that a suspect is guilty, but is excluded because it was obtained illegally. Even though there is no doubt as to the trustworthiness of that evidence, it is excluded in the interest of protecting the suspect's rights. Similarly, a potential psychological experiment may promise to provide hopeful evidence regarding the truth of some hypothesis, but may be unethical to perform because of the potential trauma it might inflict on subjects. A second consideration concerns the potential costs, in practical terms of the day-to-day use or maintenance of the evidence source. Collecting eye-witness testimony, for instance, is relatively cheap, whereas a high-tech electromagnetic spectral analyser-device is much more expensive in terms of resources and manpower. Such a device must have stronger epistemic credentials to be used widely as an evidence source.

Again, this list is not exhaustive; further, some of the categories employed (i.e. ‘costs,’ ‘moral implications,’ etc.) have an unavoidable element of vagueness to them. How should costs be weighed against each other, or against the potential benefits of using certain sources of evidence? I provide no systematic method by which to answer these questions. Here I can only hope that our pre-theoretic grasp of the notions will be enough to give meaningful content to (H*) as a potential rival to (H).

5. UNDERCUTTING WEINBERG'S CASE FOR (H)

In what follows I pursue two key goals: First, I will accommodate with (H*) the data Weinberg cites in support of (H), and second, I present data that (H) is seemingly unable to explain, but which is handily explained by (H*). In so doing, I aim to provide a cumulative case for (H*) and against (H), which will in turn undercut Weinberg's empirical case against epistemic intuitions.

Let me first turn to Weinberg's cases involving non-scientific epistemic practices such as journalism and jurisprudence. Reporters are required to have corroborating information for most of their claims; in addition, quotations from sources may only be used if they are ‘on the record.’ Weinberg contends that these features of journalism illustrate the importance of error detection in journalism (Weinberg Reference Weinberg2007: 333). The question is whether or not this commitment to error detection is best explained by a general commitment to hopefulness in any epistemic practice. To raise doubt about this latter claim, I offer the following alternative explanation: It is truistic to say that the role of a journalist is to report facts, not to engage in speculation and/or analysis of those events. The journalist is not in the business of advancing controversial claims which are then debated by the community of journalists. Thus, for a journalist to report that p is for that journalist to represent herself as being able to verify that p to the community of her peers, and therefore a journalist has a responsibility to be able to justify that degree of confidence in her claims. And performing this latter task requires that the journalist's claims be checkable in a highly reliable way. To report something one cannot corroborate is seen as irresponsible; by contrast, for a philosopher or literary critic to advance some claim she does not know is not viewed in the same way. This almost moral dimension of journalistic responsibility suggests that the requirement of corroboration is at least partly explained by moral/pragmatic considerations unique to the field of journalism itself.

In jurisprudence, Weinberg cites the appeals process as evidencing a general commitment to hopefulness in our epistemic practice (333). Here I again suggest a more pragmatic explanation in terms of commitment to the general thesis that it is far worse to convict an innocent than to allow a guilty man to go free. Because the disvalue of convicting an innocent is so great, the appeals process functions as a potential check against the permission of serious moral wrongs. If the rationale reflected a purely epistemic concern with getting things right, we would expect a similar error detection process designed as a check against wrongful acquittals. No such error detection mechanism exists. So I conclude that a commitment to hopefulness is not principally what underwrites the appeals process.

In the sciences, Weinberg cites the endorsement of electron microscopy and the collapse of Titchener's structuralist psychology as examples where the viability of an evidence source turned on the issue of hopefulness. Electron microscopy was not endorsed by the scientific community at large until its results could be corroborated by other methods. By contrast, Titchener's structuralism took as its sole source of evidence about the mind carefully performed introspections by highly trained individuals. This introspective method was entirely without hope: Introspections were judged to be accurate or inaccurate with reference to certain ‘correct’ introspections, often judged to be so by Titchener himself, and there was no independent method of corroborating his judgments (Hothersall Reference Hothersall2004: 146). The hopelessness of introspection is a large part of what led to the collapse of the movement.

Again I avert to pragmatic considerations to explain the importance of hopeful sources in the sciences. To begin with electron microscopy, electron microscopes were extremely expensive to build and maintain; from a cost-benefit point of view, the expected value of using the source would have to be very high to justify the widespread use of these devices. And the expected value is largely a function of the probability that the device delivers accurate results; thus, the device must be extensively tested for accuracy to justify its use. If widely used, the data delivered by electron microscopes could then presumably be used to fund further potentially expensive research, so the costs of error can balloon quite rapidly. Perhaps more importantly, electron microscopy has applications in medicine and virology; inaccuracies in the microscope will lead to further errors down the road in making diagnoses, devising medicines, etc. Thus the costs, in practical terms, of the device having a non-negligible chance of making errors is extremely high, and these pragmatic considerations generate incentives toward hopefulness.Footnote ⁴

Appealing to the same basic argumentative strategy I have been using thus far, I will conclude this section by turning to the case of introspection. The breakdown of Titchener's structuralism can be attributed to a variety of factors, and for my own part I am uncertain that we can definitively say that the issue ultimately turned on the hopelessness of introspection. To be sure, introspection as a scientific method was hopeless, both in the technical sense and the more general sense. The method was criticized on the grounds that because of the time delay between the actual experience and the subject's description of that experience, introspections are always retrospections (Hothersall Reference Hothersall2004: 148), and that ‘introspective reports of consciousness seem remote from consciousness as it is actually experienced. They are dull and irrelevant and certainly not of any functional value’ (148). In Reference Dunlap1912, Knight Dunlap published an influential article entitled ‘The Case against Introspection’ which denied even the reality of introspection, as it was understood by structuralists.

Yet the hopelessness of introspection as an evidence source was not the only contributing factor to the movement's demise. Titchener's system was highly restrictive, taking as its subject matter the study of the elements of a normal adult mind: Mental illness and child psychology were entirely excluded, as was any other potentially practical application of psychology as a science. These factors, combined with the emergence of competing research paradigms of psychological functionalism and behaviorism, exerted tremendous pressure on structuralism as a movement. Structuralism as a substantive psychological system threatened to stifle promising avenues of research. At the same time, attempting to separate the method of introspection from the movement and applying it to these restricted areas of research would have presented its own problems. For instance, because of the pragmatic and moral significance of mental illnesses and the diagnosis thereof, appeal to ‘hopeless’ evidence sources in this area could have very serious pragmatic and ethical implications. The demand for hopefulness is greater because the potential costs of error are also greater. These pragmatic considerations specific to the sciences seem sufficient to explain the failure of introspective evidence in psychology, yet they do not obviously generalize to disciplines outside of the sciences. So it is not yet clear that the lesson to be learned is that there is a general commitment to hopeful evidence sources, and a general rejection of all evidence that is not hopeful to a specified degree.

6. EVIDENCE FAVORING (H*) OVER (H)

In the present section I present data which prima facie favors (H*) over (H): I suggest that there are established disciplines and practices which make at least occasional appeal to ‘hopeless’ evidence sources. The first such example is jurisprudence, perhaps a surprising example given that Weinberg cites jurisprudence in presenting his case for hope. My suggestion here is that courts of law can and do frequently admit ‘hopeless’ sources into evidence for consideration by the jury. Uncorroborated witness testimony, for instance, is not automatically excluded, and the reason for this failure to exclude is just what we would expect given (H*): Such testimony is frequently our only access to any information about certain aspects of a case (i.e. no alternatives are available).

Testimony by the defendant is a particularly prominent example: The defendant is likely to be biased in his own favor, given that considerations of self-interest make it unlikely he will testify to his own guilt. Further, it is difficult to make a judgment as to the reliability of the witness in a way that is independent of the judgment of his guilt or innocence. For similar reasons, such testimony will always fail to exhibit theoretical illumination: If the suspect is innocent, the process by which he reports on the events involves perception, memory, etc. If the suspect is guilty, the process involves intentional deceit. To make a judgment about which process is in play is practically to already judge the guilt or innocence of the suspect. So it is not a stretch to suggest that testimony by the defendant is frequently ‘hopeless.’ If our judicial practice were committed to (H), such testimony should automatically be inadmissible. The fact that it is not counts against (H), and in favor of (H*).Footnote ⁵

An important possible objection on behalf of Weinberg is to argue as follows: Such testimony may be admitted as evidence in court, but it is not (or at least should not be) used single-handedly as the basis for conviction or acquittal. If a suspect is convicted solely on the basis of uncorroborated eyewitness testimony and no other evidence, then plausibly, something has gone wrong. But in much analytic philosophy, intuitions are often used to singlehandedly dismiss theories: If a theory is found to be inconsistent with a given intuition such as the Gettier intuition, then that theory is rejected. This practice is analogous to convicting solely on the basis of uncorroborated testimony or otherwise ‘hopeless’ evidence. And it is exactly this practice which Weinberg seeks to criticize.

Two things may be said in response: The first is that unlike jurisprudence, there are gradations of endorsement in philosophy other than ‘beyond reasonable doubt.’ For instance, one might claim that an intuition renders a thesis prima facie plausible, or generates a presumption in its favor, whereas in jurisprudence, there is little sense speaking of evidence as generating a prima facie presumption in favor of guilt (there is a presumption of innocence, but this is for wholly non-epistemic reasons). Intuitions are generally used to substantiate conclusions weaker than outright guilt or innocence. So the analogy between intuitions and conviction on the basis of one ‘hopeless’ evidence source breaks down.

The second is that it would be uncharitable to describe analytic philosophers' practice as instantaneously endorsing or rejecting theses solely on the basis of intuition. The practice which, according to Weinberg, analytic philosophers follow is not the one they actually employ. In some cases, such as the Gettier case, something similar to it happens, particularly when the intuition commands widespread endorsement by the community of scholars. In other cases, intuitions are not the sole court of appeal. An intuition may be part of a cumulative case against a theory, for instance. While I have no survey data to substantiate this claim, I would suggest that philosophers seldom have just one argument against views they disagree with. Further, intuitions get checked against other intuitions for coherence, are examined for possible sources of error due to various biases, etc. Under critical scrutiny, intuitions can and do get rejected as spurious. While such ‘checking’ of intuitions may not rise to the level of rendering the practice hopeful, they again suggest that the above criticism of intuitions is unsuccessful.

Beyond jurisprudence, I suggest we may look at a variety of other practices where ‘hopeless’ evidence is employed. Some useful examples here include casting for film/TV/theatre, scouting potential recruits for various forms of athletics, etc. I suggest that these practices, while far from perfect in their present form, should at least not be abandoned wholesale due to the lack of relevant alternatives. There are no purely algorithmic or scientific methods of determining who is best for a role in a movie, or what college athlete will most succeed at the professional level, etc. Yet at the same time, all of these practices crucially involve the appeal to a ‘hopeless’ source of evidence: The considered judgments of reviewers/casting directors/scouts.Footnote ⁶

None of these sources exhibit internal coherence: Scouts routinely disagree about which athletes will become the best professionals, with no apparent way to resolve the disagreement; casting directors can frequently disagree about whether or not a certain actor or actress is right for a role, etc. There is no method of external corroboration in any of these practices: There is no independent method of determining the quality of an athlete or actor, and checking to see if the judgments of casting directors or scouts track the verdicts of that independent method. Error detection is generally ex post facto, if at all: If we see an actor give a weak performance on film, we know he was miscast, but this information is not always available during casting. Much the same is true of scouting and recruiting for sports teams, as the litany of ‘draft busts’ in various sports suggests. As regards the degree to which these practices are theoretically illuminated, a detailed discussion of the psychological literature is not possible here; I will only say that I am doubtful that our understanding of the mechanisms in play is sufficient to allow these practices to cross the threshold of hopefulness while epistemic intuition remains left out.

I fully grant that none of these practices are on the same level of epistemic standing as any of the sciences, journalism, jurisprudence and the like. But it is one thing to claim this, and another to claim that the practices of casting, scouting and peer review ought to be abandoned wholesale. This latter claim strikes me as implausible and impractical, given that we have seemingly no alternatives to these methods, and given that decisions on such matters must be made within some reasonable amount of time. Yet because of its commitment to hopefulness, (H) would require us to abandon these practices whether or not we have any alternatives available. (H*), on the other hand, would permit us to make use of ‘hopeless’ evidence sources in these cases, because all of the relevant conditions are met. No alternatives are available, and there are no pragmatic considerations particular to these disciplines that rule out such evidence. I take this difference between (H) and (H*) to be a strong reason to favor (H*) over (H).

7. USING (H*) TO VINDICATE INTUITIONS

I wish to close with some brief thoughts regarding the application of principle (H*) to the use of intuitions. For all that has been said thus far, it may turn out to be the case that (H*) permits the use of ‘hopeless’ evidence sources in certain contexts, but that the standard philosophical programs of understanding knowledge, identity, morality, etc. are not among these contexts. If this were to turn out correct, then the attempt to undermine Weinberg's critique of intuitions by appeal to the superior explanatory power of (H*) would be a non-starter. So here I will provide some prima facie considerations in favor of thinking that intuitions satisfy the conditions relevant for the permissibility of using ‘hopeless’ evidence sources, as (H*) would have us conceive of those conditions.

The first condition specified by (H*) is this: A hopeful evidence source may only be used if there are no alternative evidence sources available. That this is the case with respect to intuitions about a variety of philosophical questions Weinberg seems willing enough to concede, and I find his diagnosis of the situation plausible:

The second condition that would have to be met is this: There are no pragmatic considerations specific to the relevant discipline [i.e. philosophy] that would prohibit the use of the evidence source. What sorts of pragmatic considerations would count against the use of intuitions in philosophy? I cannot attempt to consider anything like an exhaustive list of relevant considerations, so I will confine myself to what I take to be the two most important ones: the pragmatic costs of simply making use of the evidence source, and the pragmatic costs of making errors. To begin with the former, the costs of using intuitions is not prohibitively high. Compared with an expensive piece of equipment like an electron microscope, intuitions seem rather cheap: All they require is an armchair (unless you prefer to do your thinking while pacing vigorously) and the willingness to think long and hard about counterfactual scenarios. A second potential pragmatic concern revolves around the potential costs of error: Could an erroneous intuition lead to grave negative consequences for philosophical theorizing or even negative consequences outside of philosophy?

I suggest that the answer to this question is ‘no.’ Typically, if I were to endorse a false thesis on the basis of intuitions, and then get a paper published on the basis of said intuitions, this outcome would not be particularly worrisome. It is in the first place unlikely that my paper will receive general acceptance among the philosophical community at large, resulting in the whole community theorizing on the basis of my error. Even if by chance my paper does receive such acceptance, it is unclear that this acceptance would be for any extended period of time. It is often noted that if p is generally accepted by philosophers, the next wave of dissertations will be dedicated to arguing that not-p. Further, even if neither of these two scenarios is realized, and my error is accepted for an extended period by the majority of philosophers, this may still have results that are beneficial in the long run, as further refinements and reactions to my thesis may lead to better theorizing down the road. Errors in philosophy can be and frequently are beneficial, and unlike in many of the hard sciences, the costs associated with acquiring these benefits seem fairly low.

There are at least two legitimate concerns about the above response to this objection. The first is that in arguing that intuitions satisfy the conditions specified in (H*), I have effectively argued that philosophy is somehow trivial or unimportant. In response, I would distinguish two things: (A): The importance of having intelligent and well-trained people pursue high-level inquiry into matters of existential and philosophical importance; and (B): the importance of some particular philosophers making an error based on an appeal to an erroneous intuition. I have argued that (B) is not a matter on which a great deal hangs, and that the cost of errors like that described in (B) is very low both to philosophy and to those outside the discipline. Neither of these claims in any way entails that the activity described in (A) is in any sense unimportant. Consider the following argument in the context of literary theory:

1. 1. Bill Shakeshaft defended a false interpretation of Hamlet in his most recent book.

2. 2. Bill's false interpretation of Hamlet did not have any negative consequences for other literary theorists, or for people outside of literary theory.

3. 3. Therefore, literary theory is unimportant.

The inference from (1) to (2) is plainly invalid. Further, there seems no particularly plausible way to fill it out as an enthymeme. Similarly then, my claim that the costs of making errors as a result of intuitions is not particularly high does not imply that the practice of philosophy itself is a trivial one.

The second concern with my response is that there is a significant disanalogy between me attempting to get a philosophical paper which relies heavily on intuitions published and the kinds of cases where intuitions are likely to have significant consequences. Namely, I am not publishing landmark work on the level of Rawls, Singer, Judith Jarvis Thomson or any other notable figure in philosophy, let alone the great philosophers throughout the history of the discipline. There are a small number of philosophers who produce work that is of outsized importance, and this work can have political and social impact outside of what the community of philosophers does in the relative safety of academia. Does this not then raise the possibility that there could be a work that (a) is rooted heavily in intuitions, and (b) does have serious negative implications for society as a whole? If there are cases satisfying (a) and (b), this would undermine my defense of the claim that intuitions satisfy conditions (i) and (ii) as specified in rule (H*).

I will consider two possible responses to this objection. The first response is concessive in character, and begins with the empirical assumption that some philosophical views are, if widely endorsed or acted upon, more likely to have potentially far-reaching concrete consequences than others. For ease of exposition, let us assume that it is moral, social and/or political theories which fall into this category.Footnote ⁷ On the concessive view, we ought to follow the experimental critic of intuitions when it comes to intuitions about moral and political matters. Because the stakes are higher, condition (ii) of principle (H*) is no longer met. Many epistemic and metaphysical intuitions (i.e. the Gettier case, intuitions about material constitution, etc.) would remain relatively unscathed. Although this response seems plausible to me as far as it goes, it will not be satisfying for many. Thus, I will conclude with a consideration of some more encompassing responses to this objection.

The second response begins with the observation that the mere possibility of any bad consequences, even seriously bad ones, to the use of an evidence source is not automatically enough to justify abstaining from its use altogether. This is evident from some of the examples of ‘hopeless’ evidence in jurisprudence; for example, the defendant's testimony may be so emotionally affecting on a jury that it allows a dangerous murderer to go free. Testimony admitted under hearsay exception rules may also turn out to be misleading. What needs to be considered is the likelihood of these bad consequences coming about as well as how serious the consequences would be if they did indeed come about. In the case of electron microscopy, if it is unreliable, then it is highly likely that the outcomes will be both serious and widespread. So the question becomes how high the probability of negative outcomes attributable to the use of a ‘hopeless’ evidence source must be, and how serious they must be, before that evidence source may be rejected.

There will necessarily be a great deal of vagueness here. However, I am reasonably confident that the typical use of intuitions in philosophy falls well below that threshold. On the instances where a given philosophical view does indeed have grave results (although Nietzsche is not exactly an intuition-driven philosopher, let us use the appropriation of Nietzsche's ideas by the Nazi and fascist parties of Germany and Italy as an example), we have to ask whether or not these grave consequences are actually attributable to the use of intuitions, or ‘hopeless’ evidence sources more generally. Did members of the Nazi party begin from a position of relative neutrality only to find themselves persuaded by (their interpretation of) the arguments of Nietzsche, or did they begin with a certain ideology in mind and use Nietzschean ideas as a post hoc rationalization of that ideology? If the answer is closer to the latter, then it is not at all clear that a prohibition on the use of intuitions in philosophy would have done anything to limit the horrors of the Third Reich.

The kind of scenario that needs to be a non-negligible possibility in order to justify the present objection to my arguments is one in which (a) some philosophical thesis is widely adopted and the consequences are unambiguously bad,Footnote ⁸ and (b) the adoption of this thesis is attributable at least in part due to the use of intuitions to justify it. A comprehensive empirical study is obviously not possible here, but I will venture to say that the chances of (a) and (b) occurring together are negligible. Given this, the ratio of instances where philosophical theses are justified by appeal to intuitions as compared with instances of intuition use in general seems low enough, prima facie, to justify supposing that the typical use of intuitions in philosophy falls below the threshold necessary to consider them too ‘hopeless’ for any kind of use.Footnote ⁹